Information fusion technology based train wheel set tread damage online detection and recognition method

Abstract

An information fusion technology based train wheel set tread damage online detection and recognition method comprises the steps of 1 image acquisition, 2 image pre-processing, 3 preliminary damage detection, 4 accurate damage positioning and 5 damage judgment, wherein the step 4 is that an infrared image processing result and a gray image processing result are subjected to data fusion, a tread image is marked and positioned, damaged texture and area are obtained through morphological processing, morphological processing is conducted on tread cracks to obtain the length of the cracks, and the step 5 is that features of a tread damage region are extracted and selected, a BP neural network is designed and trained, the defects of the BP neural network are classified, and reference is provided for operation situation analysis of wheels, locomotives and lines according to train wheel set tread damage information recorded by a terminal processing unit. The information fusion technology based train wheel set tread damage online detection and recognition method has the advantages of being comprehensive in detection, high in automatic degree, high in detection speed and detection accuracy and the like.

Description

technical field [0001] The invention belongs to the technical field of optical measurement, and in particular relates to an online detection and identification method for tread damage of a train wheel set. Background technique [0002] The wheel is an important out-of-shape part of the train, which plays a key role in the safe operation of the train. The operating environment of train wheels is extremely harsh. In addition to impact and friction with rails and turnouts, they will also be eroded by rain, oil, and corrosive agents. In this way, wheels will often be damaged and missing to a certain extent after running for a period of time. Especially when the train is in emergency braking, the high temperature due to friction will cause the wheel tread to be scratched, peeled off, cracked and lost, and these phenomena are collectively referred to as tread damage. Train tread damage is the most difficult problem in the online detection of train wheelsets. A parameter to measur...

AI Technical Summary

Problems solved by technology

[0004] However, the current detection scheme is mainly aimed at the abrasion and peeling of the wheel-set tread, and there are few online detection technologies and system reports for wheel-set tread rolling, out-of-roundness, and mixture of abrasion and peeling; The detection and identification methods are different. Due to the high-speed and high-precision requirements of online wheel-tread damage detection, the detection and identification methods of high-speed train wheel-tread damage face the following main problems:

In high-speed situations, the detection device and system may move or rotate at a small angle with the vibration of the train, and factors such as site environment, illumination changes, and weather may introduce noise, blurred bands, and shadow areas to the image, making the general detection method and technologies susceptible to disruption

[0006] (2) Wheel tread damage includes tread abrasion, tread peeling, tread rolling, out-of-roundness (local depression), etc. Before the wheel tread damage is formed, there will be cracks on the tread surface, and the tread surface will often be disturbed by factors such as oil stains and dust , the texture characteristics of these parameters are not the same, so a single detection device and method cannot effectively identify

Due to the speed requirements of the online detection and recognition method of train wheel tread damage, the general target detection and recognition method cannot meet the requirements of online high speed and high precision at the same time

Images